Plug and play: a general approach for the fusion of sensor and machine learning models
Regularized or Bayesian inversion has revolutionized our ability to reconstruct images from incomplete data.
97 MATHEMATICS AND COMPUTING↗
SEARCH · Engineering Papers
Results for “Bayesian approach”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
System risk quantification and decision making support using functional modeling and dynamic Bayesian network
Risk-informed decision-making requires a probabilistic assessment of the likelihood of success of control action, given the system status. This paper presents a systematic state transition modeling approach integrating dynamic probabilistic risk assessment with a decision-making process using a dynamic Bayesian network (DBN) coupled with functional modeling. A functional model designed with multilevel flow modeling (MFM) technique was used to build a system state structure inferred by energy, mass, and information flow so that one can verify the developed model with respect to system functionality. The MFM model represents the causal relationship among the nodes, which captures the structure of process parameters and control units. Each node may have multiple possible states, and the DBN structured by the MFM model represents the time-domain transitions among the defined states. Furthermore, the MFM-DBN integrated state transition modeling is a white-box approach that allows one to draw the system's risk profile by updating the system states and supports the decisions probabilistically with physical inference. An example of a simple heating system has been used to illustrate this process, including decision-making support based on quantitative risk profile. For demonstrating its applicability to a complex system operational decision making, a case study of station blackout accident scenario leading to the seal loss of coolant accident in a nuclear power plant is presented. The proposed approach effectively provided the risk profile along time for each option so that the operators can make the best decision, which minimizes the plant risk.
42 ENGINEERING↗
Demonstration of machine-learning-enhanced Bayesian quantum state estimation
Machine learning (ML) has found broad applicability in quantum information science in topics as diverse as experimental design, state classification, and even studies on quantum foundations. Here, we experimentally realize an approach for defining custom prior distributions that are automatically tuned using ML for Bayesian quantum state estimation methods that generally better conform to the physical properties of the underlying system than standard fixed prior distributions. Previously, researchers have looked to Bayesian quantum state tomography for advantages like uncertainty quantification, the return of reliable estimates under any measurement condition, and minimal mean-squared error. However, practical challenges related to long computation times and conceptual issues concerning how to incorporate prior knowledge most suitably can overshadow these benefits. Using both simulated and experimental measurement results, we demonstrate that ML-defined prior distributions reduce net convergence times and provide a natural way to incorporate both implicit and explicit information directly into the prior distribution. These results constitute a promising path toward practical implementations of Bayesian quantum state tomography.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Ultrafast current imaging by Bayesian inversion
Spectroscopic measurements of current–voltage curves in scanning probe microscopy is the earliest and one of the most common methods for characterizing local energy-dependent electronic properties, providing insight into superconductive, semiconductor, and memristive behaviors. However, the quasistatic nature of these measurements renders them extremely slow. Here, we demonstrate a fundamentally new approach for dynamic spectroscopic current imaging via full information capture and Bayesian inference. This general-mode I–V method allows three orders of magnitude faster measurement rates than presently possible. The technique is demonstrated by acquiring I–V curves in ferroelectric nanocapacitors, yielding >100,000 I–V curves in <20 min. This allows detection of switching currents in the nanoscale capacitors, as well as determination of the dielectric constant. These experiments show the potential for the use of full information capture and Bayesian inference toward extracting physics from rapid I–V measurements, and can be used for transport measurements in both atomic force and scanning tunneling microscopy.
36 MATERIALS SCIENCE↗
Probabilistic Photometric Redshifts in the Era of Petascale Astronomy
With the growth of large photometric surveys, accurately estimating photometric redshifts, preferably as a probability density function (PDF), and fully understanding the implicit systematic uncertainties in this process has become increasingly important. These surveys are expected to obtain images of billions of distinct galaxies. As a result, storing and analyzing all of these photometric redshift PDFs will be non-trivial, and this challenge becomes even more severe if a survey plans to compute and store multiple different PDFs. In this thesis, we have developed an end-to-end framework that will compute accurate and robust photometric redshift PDFs for massive data sets by using two new, state-of-the-art machine learning techniques that are based on a random forest and a random atlas, respectively. By using data from several photometric surveys, we demonstrate the applicability of these new techniques, and we demonstrate that our new approach is among the best techniques currently available. We also show how different techniques can be combined by using novel Bayesian techniques to improve the photometric redshift precision to unprecedented levels while also presenting new approaches to better identify outliers. In addition, our framework provides supplementary information regarding the data being analyzed, including unbiased estimates of the accuracy of the technique without resorting to a validation data set, identification of poor photometric redshift areas within the parameter space occupied by the spectroscopic training data, and a quantification of the relative importance of the variables used during the estimation process. Furthermore, we present a new approach to represent and store photometric redshift PDFs by using a sparse representation with outstanding compression and reconstruction capabilities. We also demonstrate how this framework can also be directly incorporated into cosmological analyses. The new techniques presented in this thesis are crucial to enable the development of precision cosmology in the era of petascale astronomical surveys.
79 ASTRONOMY AND ASTROPHYSICS↗
A Correct Flat Field Model For DARHT
The forward modeling approach has been utilized to obtain the density distribution during dynamic radiographic experiments using the Bayesian Inference Engine (BIE) at the Dual Axis Radiographic Hydrodynamic Test Facility (DARHT) for over two decades. The forward modeling approach attempts to incorporate all of the known physics to obtain the transmission, from an unknown density field. The density field is then optimized in an attempt to match the flat-fielded radiographic image using a Chi-squared comparison. A depiction of the forward modeling approach implemented in the BIE is given.
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
Comparison of source-location algorithms for atmospheric samplers
Numerous algorithms have been developed to determine the source characteristics for an atmospheric release, e.g., (Bieringer et al., 2017). This study compares three models that have been applied to the data collected by the International Monitoring System operated by the Comprehensive Nuclear-Test-Ban Treaty Preparatory Commission to estimate source event parameters. Each model uses a different approach to estimate the parameters. A deterministic model uses a possible source region (PSR) approach (Ringbom et al., 2014) that is based on the correlation between predicted and measured sample values. An unnamed model developed at Pacific Northwest National Laboratory (herein called BAYEST) uses a Bayesian formulation (Eslinger et al., 2019; Eslinger et al., 2020; Eslinger and Schrom, 2016). The FREAR model uses a different Bayesian formulation (De Meutter and Hoffman, 2020; De Meutter et al., 2021; Pieter De Meutter, 2021). The performance of the three source-location models is evaluated with 100 synthetic release cases for the single xenon isotope, 133 Xe. The release cases resulted in detections in a fictitious network with 120 noble gas samplers. All three source-location models use the same sampling data. The two Bayesian models yield more accurate location estimates than the deterministic PSR model, with FREAR having slightly better location performance than BAYEST. Samplers with collection periods of 3, 6, 8, 12, and 24-h were used. Results from BAYEST show that location accuracy improves with each reduction in sample collection length. The BAYEST model is slightly better for estimating the start time of the release. The PSR model has about the same spread in start times as the FREAR model, but the PSR results have a better average start time. The Bayesian source-location algorithms give more accurate results than the PSR approach, and provide release magnitude estimates, while the base PSR model does not estimate the release magnitude. This investigation demonstrates that a reasonably dense sampling grid will sometimes yield poor location and time estimates regardless of the model. The poor estimates generally coincide with cases where there is a much larger distance between the release point and the first detecting sampler than the average sampler spacing.
98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
Neural network ensembles and uncertainty estimation for predictions of inelastic mechanical deformation using a finite element method-neural network approach
The finite element method (FEM) is widely used to simulate a variety of physics phenomena. Approaches that integrate FEM with neural networks (NNs) are typically leveraged as an alternative to conducting expensive FEM simulations in order to reduce the computational cost without significantly sacrificing accuracy. However, these methods can produce biased predictions that deviate from those obtained with FEM, since these hybrid FEM-NN approaches rely on approximations trained using physically relevant quantities. In this work, an uncertainty estimation framework is introduced that leverages ensembles of Bayesian neural networks to produce diverse sets of predictions using a hybrid FEM-NN approach that approximates internal forces on a deforming solid body. The uncertainty estimator developed herein reliably infers upper bounds of bias/variance in the predictions for a wide range of interpolation and extrapolation cases using a three-element FEM-NN model of a bar undergoing plastic deformation. This proposed framework offers a powerful tool for assessing the reliability of physics-based surrogate models by establishing uncertainty estimates for predictions spanning a wide range of possible load cases.
42 ENGINEERING↗
CCUS 2024, Interpreting the strain tensor Larry Murdoch Interpreting strain tensor data to characterize and monitor reservoirs for CO2 storage and other applications
Recent advances in instrumentation have made it feasible to measure the transient strain tensor caused by small changes in fluid volume or pressure in the subsurface and this has opened the door to new opportunities for characterization and monitoring during CCUS. We have demonstrated this method by deploying strainmeters at shallow depths (30 to 40m) and then conducting injection well tests in an underlying reservoir at 530m depth. The resulting data indicated that the horizontal strain at shallow strainmeters was tensile and the vertical strain was compressive. The radial strain was less than the horizontal strain, and the strain rates decreased from 100 nanostrain/day to roughly 10 ne/d over a few days (1 nanostrain = 1 part per billion strain). We then used the strain data to estimate reservoir properties, geometry and pressure through inversion of poroelastic forward models using both numerical and novel analytical methods. The average horizontal strain in the caprock resembles the transient pressure in the underlying reservoir and classic type-curve methods from transient well testing can be used for preliminary interpretations of strain data. We have developed fast, closed-form analytical solutions to a pressurized poroelastic inclusion and inhomogeneity in a half-space. Numerical models developed using finite element methods allow more details of the subsurface to be included in the inversion, but they require much longer run times and this makes inversion cumbersome using standard methods. We have developed an inversion approach that uses a proxy model created using machine learning to do most of the forward calculations. This approach markedly reduces the computational requirements and makes it feasible to use Bayesian inversion with large numerical models. Bayesian inversion is important because it provides predictions with uncertainties, which makes the results useful for decision making. We have shown with field tests and simulations that the strain tensor in the caprock is sensitive to pressure in the reservoir, reservoir properties and boundaries, and pressure in the caprock caused by leaks. These results indicate that measuring and interpreting the shallow strain tensor could be a valuable tool for both initial reservoir characterization efforts and long-term monitoring during CCUS. Recent advances in instrumentation have made it feasible to measure the transient strain tensor caused by small changes in fluid volume or pressure in the subsurface and our objective was to evaluate opportunities for strain monitoring during characterization and monitoring for CCUS. Our approach was to deploy strainmeters at shallow depths (30 to 40m) and then conduct injection well tests in an underlying reservoir at 530m depth. The results indicate that the horizontal strain at shallow strainmeters was tensile and the vertical strain was compressive. The radial strain was less than the horizontal strain, and the strain rates decreased from 100 nanostrain/day to roughly 10 ne/d over a few days (1 nanostrain = 1 part per billion strain). We then used the strain data to estimate reservoir properties, geometry and pressure through inversion of poroelastic forward models using both numerical and novel analytical methods. The average horizontal strain in the caprock resembles the transient pressure in the underlying reservoir and classic type-curve methods from transient well testing can be used for preliminary interpretations of strain data. We have developed fast, closed-form analytical solutions to a pressurized poroelastic inclusion and inhomogeneity in a half-space. Numerical models developed using finite element methods allow more details of the subsurface to be included in the inversion, but they require much longer run times and this makes inversion cumbersome using standard methods. We have developed an inversion approach that uses a proxy model created using machine learning to do most of the forward calculations. This approach markedly reduces the computational requirements and makes it feasible to use Bayesian inversion with large numerical models. Bayesian inversion is important because it provides predictions with uncertainties, which makes the results useful for decision making. In conclusion, we have shown with field tests and simulations that the strain tensor in the caprock is sensitive to pressure in the reservoir, reservoir properties and boundaries, and pressure in the caprock caused by leaks. These results indicate that measuring and interpreting the shallow strain tensor could be a valuable tool for both initial reservoir characterization efforts and long-term monitoring during CCUS.
Murdoch, Larry↗
Nonlinear Attitude Filtering Methods
This paper provides a survey of modern nonlinear filtering methods for attitude estimation. Early applications relied mostly on the extended Kalman filter for attitude estimation. Since these applications, several new approaches have been developed that have proven to be superior to the extended Kalman filter. Several of these approaches maintain the basic structure of the extended Kalman filter, but employ various modifications in order to provide better convergence or improve other performance characteristics. Examples of such approaches include: filter QUEST, extended QUEST, the super-iterated extended Kalman filter, the interlaced extended Kalman filter, and the second-order Kalman filter. Filters that propagate and update a discrete set of sigma points rather than using linearized equations for the mean and covariance are also reviewed. A two-step approach is discussed with a first-step state that linearizes the measurement model and an iterative second step to recover the desired attitude states. These approaches are all based on the Gaussian assumption that the probability density function is adequately specified by its mean and covariance. Other approaches that do not require this assumption are reviewed, including particle filters and a Bayesian filter based on a non-Gaussian, finite-parameter probability density function on SO(3). Finally, the predictive filter, nonlinear observers and adaptive approaches are shown. The strengths and weaknesses of the various approaches are discussed.
F Landis Markley↗
Exploratory analysis of machine learning techniques in the Nevada geothermal play fairway analysis
Play fairway analysis (PFA) is commonly used to generate geothermal potential maps and guide exploration studies, with a particular focus on locating and characterizing blind geothermal systems. This study evaluates the application of machine learning techniques to PFA in the Great Basin region of Nevada. Following the evaluation of various techniques, we identified two approaches to PFA that produced promising results, 1) supervised Bayesian probabilistic neural networks to generate geothermal potential maps with confidence intervals, and 2) unsupervised principal component analysis paired with k-means clustering to generate both cluster maps to help identify spatial patterns, as well as new combined feature inputs. We applied these techniques to perform a comparative analysis between two principal sets of geological and geophysical features related to permeability and heat and a set of positive (known geothermal resources) and negative training sites (known drill sites with unsuitable geothermal conditions). We found that these methods constrain previously unrecognized feature controls on geothermal favorability, many of which are spatially organized within the extent of cluster groups and the major structural-hydrologic domains of the study area. Furthermore, we utilized exploratory unsupervised modeling to highlight spatial relationships between input data and predictive output results of our supervised modeling. As a result, we demonstrate how our models compare to the previous Nevada PFA and how the rapid insights these machine learning techniques offer may support future assessments of both known and undiscovered blind geothermal systems in the Great Basin region of Nevada and beyond.
15 GEOTHERMAL ENERGY↗
Customized Bayesian optimization for efficient beam tuning at the facility for rare isotope beams
Bayesian optimization (BO) has recently emerged as a powerful approach for on-line beam tuning, and it is rapidly gaining adoption across accelerator facilities due to its flexibility and efficiency in handling complex optimization tasks. At the Facility for Rare Isotope Beams, rapid and reliable tuning is essential to support the delivery of diverse ion species. To improve the practicality of BO in this setting, we implemented several enhancements, including scalarized composite objective construction for multicriteria optimization, asynchronous evaluation for better resource utilization, prior-mean-assisted optimization to accelerate convergence, and a local search strategy for rapid completion of the task. We present the details of these methods, discuss challenges-encountered, and share our experience applying them to specific beam-tuning tasks.
Accelerators & storage rings↗
GP Cosmology Surrogate v1.0
GP Cosmology Surrogate is a Python library for building and training a generalized multi-output Gaussian process (GP) framework of @takhtaganov2021cosmic. In this approach, the surrogate is constructed sequentially, guided by a Bayesian optimization acquisition function that targets reduction of emulation error in the regions most consistent with the observational data. This adaptive design concentrates computational resources where they have the greatest impact on inference accuracy. The library supports efficient training for separable GP kernels, which allows the use of Kronecker algebra to handle high-dimensional input spaces and large numbers of correlated outputs. This makes it well suited for applications such as modeling cosmological power spectra, large-scale physical simulations, and multi-output hyperparameter tuning. By combining scalable multi-output GP modeling with data-driven adaptive sampling, GPsurrogate enables parameter inference and optimization with substantially fewer simulations than conventional space-filling designs.
Lukic, Zarija [Lawrence Berkeley National Laborato↗
Overview of the LANL Multi-Probe Radiography Project [Slides]
Future radiographic facilities for the U.S. defense program will be required to provide more information as simulation codes improve in both physics’ fidelity and resolution. A possible approach is to use more types of probe beams in addition to, or instead of X-rays, generated by 20-MeV electron accelerators from a small number of directions. High power short-pulse laser systems can generate beams of protons, neutrons and electrons, as well as X-rays. The cost of these systems is falling rapidly. So, it can be imagined that deploying multiple short-pulse lasers along with other, more traditional probes, will become feasible. In this project, we are following three paths to determine if such an approach will succeed for cm-scale objects. The first is an experimental one to determine if the presence of multiple short-pulse probes cause interference with each other, especially while radiographing dynamic objects. The second leg of this project is to determine if having multiple types of probes really does give more information on composition. Finally, an overall assessment of the viability of this approach will be made. Examples from recent experiments at the Omega EP laser will be presented. The initial approach to evaluating radiographs with multiple probes using the Bayesian Inference Engine (BIE) also will be given.
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
ANALYZING METEOROID FLIGHTS USING PARTICLE FILTERS
Fireball observations from camera networks provide position and time information along the trajectory of a meteoroid that is transiting our atmosphere. The complete dynamical state of the meteoroid at each measured time can be estimated using Bayesian filtering techniques. A particle filter is a novel approach to modeling the uncertainty in meteoroid trajectories and incorporates errors in initial parameters, the dynamical model used, and observed position measurements. Unlike other stochastic approaches, a particle filter does not require predefined values for initial conditions or unobservable trajectory parameters. The Bunburra Rockhole fireball, observed by the Australian Desert Fireball Network (DFN) in 2007, is used to determine the effectiveness of a particle filter for use in fireball trajectory modeling. The final mass is determined to be 2.16±1.33 kg with a final velocity of 6030±216 m s{sup −1}, similar to previously calculated values. The full automatability of this approach will allow an unbiased evaluation of all events observed by the DFN and lead to a better understanding of the dynamical state and size frequency distribution of asteroid and cometary debris in the inner solar system.
79 ASTRONOMY AND ASTROPHYSICS↗
Deep Galex Observations of the Coma Cluster: Source Catalog and Galaxy Counts
We present a source catalog from deep 26 ks GALEX observations of the Coma cluster in the far-UV (FUV; 1530 Angstroms) and near-UV (NUV; 2310 Angstroms) wavebands. The observed field is centered 0.9 deg. (1.6 Mpc) south-west of the Coma core, and has full optical photometric coverage by SDSS and spectroscopic coverage to r-21. The catalog consists of 9700 galaxies with GALEX and SDSS photometry, including 242 spectroscopically-confirmed Coma member galaxies that range from giant spirals and elliptical galaxies to dwarf irregular and early-type galaxies. The full multi-wavelength catalog (cluster plus background galaxies) is 80% complete to NUV=23 and FUV=23.5, and has a limiting depth at NUV=24.5 and FUV=25.0 which corresponds to a star formation rate of 10(exp -3) solar mass yr(sup -1) at the distance of Coma. The GALEX images presented here are very deep and include detections of many resolved cluster members superposed on a dense field of unresolved background galaxies. This required a two-fold approach to generating a source catalog: we used a Bayesian deblending algorithm to measure faint and compact sources (using SDSS coordinates as a position prior), and used the GALEX pipeline catalog for bright and/or extended objects. We performed simulations to assess the importance of systematic effects (e.g. object blends, source confusion, Eddington Bias) that influence source detection and photometry when using both methods. The Bayesian deblending method roughly doubles the number of source detections and provides reliable photometry to a few magnitudes deeper than the GALEX pipeline catalog. This method is also free from source confusion over the UV magnitude range studied here: conversely, we estimate that the GALEX pipeline catalogs are confusion limited at NUV approximately 23 and FUV approximately 24. We have measured the total UV galaxy counts using our catalog and report a 50% excess of counts across FUV=22-23.5 and NUV=21.5-23 relative to previous GALEX measurements, which is not attributed to cluster member galaxies. Our galaxy counts are a better match to deeper UV counts measured with HST.
Hammer, D.↗
Integration of Condition-Based, Diagnostic, Prognostic, And Anomaly Detection Data into Reliability Models to Support a Predictive Maintenance Context
Reliability data employed in plant reliability models are an approximated integral representation of the past industrywide operational experience, and they neglect the present asset health status (available, for example, from online monitoring data and diagnostic assessments) and forecasted health projection (when available from prognostic models). Ideally, in a predictive maintenance context, system reliability models should support decision making by propagating actual health information from the asset to the system level in order to provide a quantitative snapshot of system health and identify the most critical assets. Asset health should be informed solely by that specific asset’s current and historical performance data and should not be an approximated integral representation of the past industrywide operational experience (as currently performed by system reliability models through Bayesian updating processes). This paper proposes a reliability modeling approach that relies on asset diagnostic and prognostic assessments, along with monitoring data to measure asset health. We show how state-of-the art condition-based, diagnostic, prognostic, and anomaly detection models can be linked to system reliability models not in probability terms, but in terms of margin where margin is defined as the “distance” between the present status and an undesired event (e.g., failure or unacceptable performance). Then, we show how the propagation of margin data from the asset to the system level is performed through classical reliability models such as fault trees or reliability block diagrams. The described method is in fact able to propagate heterogenous health data from the asset to the system level in order to analytically assess system health.
97 MATHEMATICS AND COMPUTING↗
Combining Poisson singular integral and total variation prior models in image restoration
In this study, a novel Bayesian image restoration method based on a combination of priors is presented. It is well known that the Total Variation(TV) image prior preserves edge structures while imposing smoothness on the solutions.However,it tends to oversmooth textured areas.To alleviate this problem we propose to combine the TV and the Poisson Singular Integral(PSI)models, which,as we will show, preserves the image textures. The PSI prior depends on a parameter that controls the shape of the filter. A study on the behavior of the filter as a function of this parameter is presented. Our restoration model utilizes a bound for the TV image model based on the majorization–minimization principle,and performs maximum a posteriori Bayesian inference. In order to assess the performance of the proposed approach,in the experimental section we compare it with other restoration methods.
97 MATHEMATICS AND COMPUTING↗